Autokers textinput text classifier does not work.

[ferdasonmez]

Hi,
I am trying to use Autokeras to classify varied length of text (which are lines from a code file).
I am hoping to get a model design thisway, however I am getting the below error. On purpose I did not tokenize or pad the data so that I am hoping the autokeras can provide the model which does that for me in a proper way.
Regards,
Ferda

File "C:\Programfiles\Python311\Lib\site-packages\autokeras\auto_model.py", line 344, in
x_shapes = [a.shape for a in nest.flatten(x)]
^^^^^^^
AttributeError: 'str' object has no attribute 'shape'

    input_node = ak.TextInput()
    output_node = ak.TextBlock(block_type="ngram")(input_node)
    output_node = ak.ClassificationHead()(output_node)
    clf = ak.AutoModel(
        inputs=input_node, outputs=output_node, overwrite=True, max_trials=1
    )

    clf.fit(X_train_text, y_train[:, 0], epochs=2,  validation_data=(X_val_text, y_val))
    # Export as a Keras Model.
    model = clf.export_model()